Question: "Does the use of face marks in the general population make a difference to spread of infection?" - Gov.uk
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Rapid review on face masks for the general population, version 7/4/2020
Question: “Does the use of face marks in the general population make a difference
to spread of infection?”
Answer
• Based on the evidence from three recent systematic reviews and meta-analyses
[including our re-analysis focusing on community trials] wearing face masks in the
community was not significantly associated with a reduction in episodes of
influenza-like illness [ILI]; the overall assessment of the quality was classified as low.
o Jefferson 2020 [re-analysed]: 7 RCTs in the general population with ILI
outcome [OR (95% CI) 0.92 (0.87, 1.07)]
o Xiao 2020: 10 RCTs in non-healthcare settings with pandemic
influenza outcomes [OR (95% CI) 0.97 (0.79, 1.18]
o Brainard 2020: various study designs with respiratory illness outcome; OR
(95% CI): 0.94 (0.75, 1.19)
• SARS-CoV-2 is transmissible by contact and droplets [aerodynamic diameter >5μm].
SARS-CoV-2 can be detectable and viable in aerosols [aerodynamic diameter ≤5μm],
suggesting possible transmission routes by aerosols. However, there is little current
evidence demonstrating actual aerosol transmission episodes by SARS-CoV-2.
• The quality of the evidence on face mask effectiveness is moderate to low. See table
1. Many of the cohort and cross-sectional studies rely on self-reported symptoms not
confirmed clinically or using lab tests. There is very little information on duration or
frequency of use or correct usage of masks.
• Whilst some of the RCTs specify the type of mask used, many of the studies do not
define the type of mask or the materials masks are made from. This makes it difficult
to evaluate the evidence.
• Mask-wearing alone, in the absence of other preventive measures, is unlikely to be
effective, yet most studies do not take this into account. Many studies did not gather
information on general hygiene and other relevant health behaviours (e.g. hand
sanitiser, hand-washing). Many of the studies do not make a distinction between
indoor and outdoor settings.
• Much of the evidence is not generalizable to a UK community setting. For example, 8
of the 24 studies focus on face mask use during the annual hajj pilgrimage in Saudi
Arabia – a very specific context in very different climatic conditions. The influence of
cultural and socio-behavioural factors (e.g. fear, stigma, altruism) on levels of
compliance during a pandemic may differ meaningfully from other circumstances.
• There is little evidence on the behavioural aspects of facemask use. The most-studied
aspect relates to frequency / consistency of use, with more consistent use linked to a
greater reported protective effect (although this must be taken in the context of our
overall findings which failed to find a clear protective effect of facemasks). One study
found that facemasks contribute to an increased sense of isolation.
• Public health awareness campaigns [Aiello-2010], specific education [Barasheed-
2016] and provision of free facemasks [Alabdeen-2005] all appeared to incentivise
greater uptake of facemasks. There were little data on how long people can be
expected to comply with requirements to wear a facemask. One review reported that
1Rapid review on face masks for the general population, version 7/4/2020
Extended abstract
Background: Current UK advice advises that “respiratory etiquette when coughing or
sneezing ” and social distancing of at least 2m apart should give sufficient protection against
transmission from viruses carried in droplets which evaporate or fall to the ground within that
distance. However, recent data has suggested that exhalation, coughing and sneezing can
carry liquid droplets / aerosols over larger distances and has led to renewed interest in the
role of facemasks to limit transmission risk. If there were a general recommendation to wear
face masks indoor when symptomatic, or outdoors in public is there evidence to suggest that
this may help slow the spread of coronavirus? Could wearing a mask be as effective as social
distancing? The WHO Expert Panel on this topic reported on 6/4/2020 that “the wide use of
masks by healthy people in the community setting is not supported by current evidence and
carries uncertainties and critical risks”. This is in contrast to US CDC who recommended the
US public wear cloth coverings in pharmacies, groceries and other public places where social
distancing is hard to maintain.
Background policy relevance
• Can the use of masks prevent transmission of SARS-COV-2?
• Do masks reduce the virus shedding in respiratory droplets and/ or aerosols?
• Is there a difference between different types of masks (eg surgical or home-made
masks)?
• Are there behavioural aspects of face mask wearing by the general population that
relate to compliance or risk taking behaviour that are relevant?
Methods:
We adapted rapid review methods outlined by the Cochrane Collaboration. To address these
broad and complex questions rapidly, we broke the review into four main inter-connected
research questions:
• sub-review 1: what is the effectiveness of face masks in preventing respiratory
transmission in the community?
• sub-review 2: what is the relative effectiveness of medical masks versus non-medical
masks or equivalent barriers?
• sub-review 3: what important behavioural aspects of wearing masks in terms of
compliance with advice and impact on risk taking behaviour can be identified?
• sub-review 4: what is known about the nature and spread of respiratory airway
particles?
Literature Search:
Preliminary groundwork: We were aware of work to produce reviews and evidence
summaries on COVID-19 -related topics. Since there is currently no register of existing
reviews we have been compiling a database of reviews from websites of partners taking part
in the WHO Evidence Collaborative and have so far identified ~170 COVID-19 evidence
reviews, including some on use of face masks. To support this rapid review, and avoid
3Rapid review on face masks for the general population, version 7/4/2020
duplication of effort, we contacted experts in the field and searched for prior reviews or
evidence summaries on facemasks to prevent transmission of infection. We identified 14
prior reviews/summaries, appraised them for scope and quality, and for this update rapid
review selected the three most recent, on-topic, and robust quality [Jefferson 2020, Brainard
2020, Xiao 2020] for updating and re-analysis.
The methods described here apply to sub-reviews 1 and 2.
We re-ran the searches reported in the three selected reviews, limited to publications from
the date onward that each of the systematic review had stopped their search. We searched
the databases used in the prior reviews (PubMed, Medline, Embase, Scopus, CENTRAL,
CINAHL) and augmented the methods by including a search for pre-prints on medRxiv. The
searches were carried out by one reviewer (MD).
From the updated search results set, we excluded publications published before 2020, from
nosocomial settings, modelling data, animal models, providing commentary but no data. All
component studies of the three systematic reviews were included in this update. There were
no language limitations as part of the search, but due to time and resource constraints, non-
English publications were not included in analysis
Sub-review 1: What is the effectiveness of face masks in preventing respiratory transmission
in the community?
Background
Community face mask use was part of successful control policies in China, South Korea and
Vietnam, but it is not possible to disentangle their separate contribution to reducing
transmission. This rapid review was carried out to establish whether there is evidence for the
use of face masks in the general population to reduce the spread of infection with SARS-COV-
2.
Methods
For a description of the underlying literature search, please see Literature search methods
section above.
From the updated search results set, we excluded publications from nosocomial settings,
modelling data, animal models, providing commentary but no data, and, due to time
constraints studies published before 2020. All component studies of the three systematic
reviews [Jefferson 2020, Brainard 2020, Xiao 2020] were included in this update.
Screening was shared between three reviewers (MG, XL, WX). Each new title, abstract and full
text was screened by one reviewer (MG). References of previous systematic reviews were
searched by two reviewers (XL, WX). No new studies meeting the inclusion criteria were
identified.
Results
4Rapid review on face masks for the general population, version 7/4/2020
• A total of 766 new results was found from the database searching, reduced to 81 after
removal of duplicates and pre-2020 publications. We excluded 72 records by
screening titles and abstracts and a further 9 at the full text screen/quality assessment
phase, leaving 0 new articles for inclusion in the final review. The key findings from
this rapid review were:
• Of the three high quality recent reviews we scrutinised in detail, two included only
RCTs [Jefferson 2020, Xiao 2020], whereas Brainard 2020 included population studies
too. We ran updated literature searches for these reviews to identify new studies. No
new studies meeting inclusion criteria were identified.
• All component studies of the three systematic reviews were included for analysis in
this update.
• Jefferson 2020 included 9 RCTs (7 in the general population and 2 in health care
workers) and reported that there was no reduction of Influenza-like illness (ILI) for
masks compared to no masks [Random effects OR (95% CI): 0.93 (0.83, 1.05)].
• We re-ran a random effects meta-analysis restricting to the 7 RCTs conducted in the
general population from Jefferson 2020 and also found no significant reduction of ILI
[OR (95% CI): 0.92 (0.87, 1.07)]. Risk of bias analysis using the Cochrane tool done by
Jefferson et al indicated that there was high or unknown risk of bias in relation to
performance, detection and reporting bias.
• Xiao 2020 evaluated environmental and personal protective measures for pandemic
influenza in non-healthcare settings. They run a fixed effect meta-analysis of 10 RCTs
of community use of face masks (with or without hand hygiene measures) and they
reported a no significant reduction of ILI [Fixed effect OR (95% CI): 0.92 (0.75, 1.12)].
We repeated the analysis using random effects meta-analysis and the result was
similar [Random effects OR (95%CI): 0.97 (0.79, 1.18)]. The study quality of the
included studies was evaluated using GRADE by Xiao et al and the overall assessment
of the quality was classified as low.
• Brainard 2020 included all study designs on facemasks and similar barriers to prevent
respiratory illness. Based on random effects meta-analyses on RCTs, they concluded
that wearing face masks can be very slightly protective against primary infection from
casual community contact, but this was not significant, and the evidence was
classified as low certainty-evidence using the Cochrane risk assessment [Random
effects OR (95% CI): 0.94 (0.75, 1.19)]. Similar were the findings for the prevention of
household infections when both infected and uninfected members wear face masks.
Conclusion
• Based on the evidence from three recent systematic reviews and meta-analyses
wearing face masks in the community is not significantly associated with a reduction
in ILI and the overall assessment of the quality was classified as low.
Key References – see separate file
Lead reviewers [with contact details of lead]
•
5Rapid review on face masks for the general population, version 7/4/2020
Sub-review 2: What is the relative effectiveness of medical masks versus non-medical
masks or equivalent barriers?
Background
This review evaluates the evidence on the effectiveness of face masks for preventing
respiratory infection in community settings.
Methods
• For a description of the underlying literature search, please see Literature search
methods section above.
• We sought published or pre-published observational or intervention studies,
investigating face masks or respirators to prevent the transmission of respiratory
viruses in community settings. Face masks could be surgical, medical, N95
respirators, homemade, improvised or repurposed (e.g. DIY masks) made of any
material. Included studies had to report a measure of respiratory virus infection
and/or its consequences (e.g. days off work, complications, hospital admission,
deaths). We excluded case series, case reports, review articles, guidelines,
discussions, regulations, debates, and commentaries. We also excluded publications
which investigated the prevention of transmission to/from clinically trained persons
in clinical settings, studies based on mathematical modelling, and studies
investigating transmission from non-humans. Due to time constraints, studies
published before 2020. All component studies of the three systematic reviews
[Jefferson 2020, Brainard 2020, Xiao 2020] were included in this update.
• Title and abstract screening was conducted by three people, each person screening
a third of the studies. A second person checked all rejected studies. Where the
second reviewer disagreed with the decision of the first reviewer, the paper was
retained for full text screening. Full text screening was again split between the three
reviewers. Data extraction and quality appraisal were conducted by a different
reviewer from the reviewer who conducted the screening. We used the following
quality assessment checklists: CASP checklist for randomised controlled trials, cohort
and case-control studies and Joanna Briggs checklists for case series and cross-
sectional studies.
Results
• We identified a total of 182 studies (107 were primary studies from the 3 key
systematic reviews and 78 were studies identified in our update search). We
rejected 125 through screening titles and abstracts and a further 32 when reviewing
full texts. Reasons for rejection at full text screen were: not meeting inclusion and
exclusion criteria (n=18), not primary studies (n=6), full text not available (n=8). We
retained 25 studies for detailed analysis and quality appraisal. Key findings were
that:
• The quality of the evidence on face mask effectiveness is moderate to low. See table
1.
6Rapid review on face masks for the general population, version 7/4/2020
• Many of the cohort and cross-sectional studies rely on self-reported symptoms not
confirmed clinically or using lab tests.
• There is very little information on duration or frequency of use or correct usage of
masks.
• Whilst some of the RCTs specify the type of mask used, many of the studies do not
define the type of mask or the materials masks are made from. This makes it difficult
to evaluate the evidence.
• Mask-wearing alone, in the absence of other preventive measures, is unlikely to be
effective, yet most studies do not take this into account. Many studies did not gather
information on general hygiene and other relevant health behaviours (e.g. hand
sanitiser, hand-washing)
• Many of the studies do not make a distinction between indoor and outdoor settings.
• Much of the evidence is not generalizable to a UK community setting. For example, 8
of the 24 studies focus on face mask use during the annual hajj pilgrimage in Saudia
Arabia – a very specific context in very different climatic conditions. Only one lack of
transferability between different populations.
• Of the seven studies of moderate quality (table 3) – i.e. the strongest evidence found
– three reported no evidence of effectiveness of face masks, whilst 4 reported some
evidence of effectiveness. However, a key consideration is the difference between
evidence of effectiveness in a controlled study and the evidence of effectiveness in
real life situations, where compliance may not be optimum.
Conclusions
• This review found mixed and low-quality evidence on the use of face masks to
prevent community transmission of respiratory illness, with much of the evidence
generated in very different contexts from the UK. A key issue is the need for better
quality research in community settings, which focuses not only on evaluating
different types of mask but also on evaluating adherence (duration and frequency of
mask use, correct procedure for putting on and removing masks). This review was
conducted very quickly, and as such has the following weaknesses: full text
screening, extracted data and quality assessment were not checked by a second
reviewer, thus introducing a risk of bias. We will continue to update and refine this
review going forward.
Key References – see separate file
Lead reviewers [with contact details of lead]
•
7Rapid review on face masks for the general population, version 7/4/2020
Sub-review 3: What evidence is there for the role of behavioural factors on the effectiveness
of face mask use in the community?
Background
We looked at behavioural factors that are linked directly to facemask use: Is the facemask
put on and taken off correctly? How often do people wear facemasks? Does this change
over time? Do the population comply with advice on their use?
Methods
• For this sub-review, we focused on the 14 prior reviews and summaries found during
the preliminary ground-work described in the literature search methods above.
• These reviews were screened by three reviewers (EMS, MP, AN) for relevance to our
sub-question (behavioural aspects of facemask use) and 11 were identified that met
our inclusion criteria. The primary studies within these reviews were then taken
forward for title & abstract, and subsequent full-text, screening.
Screening Criteria: We included studies that considered:
• How masks are used (e.g. whether people are putting them on or taking them off
safely) and whether this alters their effectiveness;
• How mask use affects other relevant protective or risk-taking behaviours;
• Whether mask use changes in the long term; and
• What behavioural interventions (e.g. training, communications) may affect mask use.
We excluded studies that considered:
• Mask use among healthcare workers or in care settings only.
Screening and Data Extraction
• 84 primary studies were identified from the reference lists of the relevant reviews. 8
studies were excluded because full-text was unavailable, and 2 because they were not
in English, by the team who retrieved the studies (RMQ, LG and YB).
• 74 studies remained to be screened. Of these, 9 were prioritised by MP for data
extraction, based on our full-text screening of the existing reviews. Data extraction
was carried out by two reviewers (MP and AN).
• Title and abstract screening was carried out by one reviewer (EMS) for the other 65
studies, based on our inclusion criteria. 30 studies were included at this stage.
Exclusions were checked by a second reviewer (MP), and one further study was
included for data extraction.
• Data extraction on these 31 studies was carried out by three reviewers (EMS, AN and
MP). 9 further studies were excluded as a result of full-text screening, principally
because they did not include any investigation of the behavioural aspects of mask use.
Quality assessment
• We carried out a quality assessment of the remaining 22 reviews based on templates
adapted from the CASP checklists for critical appraisal.
8Rapid review on face masks for the general population, version 7/4/2020
Results
The key findings from this rapid review were:
• Behavioural aspects of mask use have not been a primary focus of any study on the
effectiveness of facemasks. A small number of studies compare the effectiveness of
occasional vs regular facemask use, but these terms are not clearly defined and the
studies depend on self-reporting of compliance.
• The limited evidence base suggests that regular/consistent use of masks may be more
protective than irregular use (but within the context of a wider literature which is
inconclusive about the general protective effect of masks). However, the difference
between ‘consistent’ and ‘irregular’ use is not clearly defined in existing studies, and
is therefore of limited use in developing guidance.
• One review found that adherence to facemask use tended to drop off after five days.
Another found that adherence depended on health beliefs and perception of risk.
• Reported concerns that people may wear masks ‘incorrectly’, and therefore
ineffectively, in the community are a feature of the literature, but there do not appear
to be any studies which assess the extent to which this actually happens, nor how it
impacts on effectiveness.
• One study found that people who wore facemasks appeared to have increased
compliance with hand hygiene practices. Of concern, however, the same study found
an increased rate of respiratory infection among non-vaccinated people who wore
facemasks. The evidence is not strong enough to allow us to conclude that facemask
use encourages either protective or risk-taking behaviours, but these findings
certainly suggest that a degree of caution should be applied.
• A small number of studies found that behavioural incentives – including specific
training, public health awareness campaigns, and provision of free face masks –
encouraged uptake of masks.
• One study addressed the barriers to use of facemasks, and found that masks
contributed to a sense of isolation from others (as well as discomfort and difficulty
breathing). This study was not carried out in the context of a pandemic, with mass
distancing and ‘lockdown’, but the possible mental health implications of this finding
may require some consideration in this context.
• Most of the studies looking at the use of masks in community settings relate to very
specific contexts: schools, university halls of residence, and, most frequently, the Hajj.
The Hajj in particular is a unique, time-limited event. Care should be taken when
generalising from these studies to the community in general.
Conclusions
• There is little evidence on the behavioural aspects of facemask use, and most studies
relate to unique, defined contexts (predominantly the Hajj). The aspect most
frequently studied relates to frequency / consistency of use, and it is suggested that
more consistent use is linked with a more protective effect (although this must be
taken in the context of overall findings about the [limited] protective effect of
facemasks).
9Rapid review on face masks for the general population, version 7/4/2020
• One study found that facemasks contribute to an increased sense of isolation, while
another found higher rates of respiratory infection among some participants who
wore a facemask, which may hint at a link between facemask use and risk-taking
behaviours. Neither of these findings is supported by substantial or robust evidence,
but both might merit further research in order to inform a full appraisal of the costs
vs benefits of facemask use in community settings.
Key References – see separate file
Lead reviewers [with contact details of lead]
•
Sub-review 4: What is the mode of transmission of SARS-CoV-2 and other common
respiratory pathogens?
Background
• This rapid review was conducted to address the question of whether an
understanding of SARS-CoV-2 transmission routes can help inform decisions
regarding community use of face masks.
Methods
Two working strands were conducted in parallel to address the question.
• Strand 1 searched for original studies and reviews that reported the mode of
transmission of coronaviruses, including SARS-CoV-2, MERS-CoV, SARS-CoV-1, and
seasonal coronaviruses (i.e. NL63, 229E, OC43 and HKU1).
• Strand 2 searched for existing reviews that reported the mode of transmission of
common human respiratory pathogens.
Selection criteria are in the Appendix. As studies applied different approaches to infer mode
of transmission, we grouped the approaches into three levels based on the strength of the
evidence:
• Level 1. Pathogen being detectable (in aerosols, droplets or surfaces);
• Level 2. Pathogen being detectable and viable;
• Level 3. Actual transmission events being confirmed. All studies were extracted to an
extraction template attached in Appendix.2.
Results
A total of 25 studies were included and their findings were summarised in Table 1. Key
findings include:
• All respiratory pathogens included in the review can be transmitted by
direct/indirect contact and droplets.
• Measles, influenza virus and adenovirus are known to be transmissible by aerosols.
10Rapid review on face masks for the general population, version 7/4/2020
• SARS-CoV-2 can be detected and is viable in aerosols but with no direct evidence of
transmission via aerosols.
Conclusions
• SARS-CoV-2 is transmissible by contact and droplets.
• SARS-CoV-2 can be detectable and viable in aerosols, suggesting possible transmission
routes by aerosols. However, little evidence is available so far demonstrating actual
aerosol transmission episode by SARS-CoV-2.
Key References – see separate file
Lead reviewers [with contact details of lead]
•
Table 2. Summary of findings on mode of transmission of common human respiratory
pathogens [see separate file]
11You can also read
